Physical Bases of Cognitive Flexibility

Abstract

Cognitive flexibility-the ability to rapidly adapt our thoughts and behavior-is the foundation of intelligence in challenging environments. However, flexibility is also easily compromised in high-stress environments, often by processes that limit metabolic energy. Understanding the metabolic costs of cognitive flexibility could be the key to making cognitive flexibility more robust in challenging environments. Building on a powerful new computational method that can identify periods of cognitive flexibility, the research objective of this proposal is to test the hypothesis that cognitive flexibility depends on neural computations that are unusually energetically costly. Our technical approach includes 3 mutually-reinforcing aims in humans and non-human primates, in whom we can directly measure the neural computations involved in cognitive flexibility. Aim 1 simultaneously characterizes the energetic costs and computational demands of cognitive flexibility in the non-human primate. It combines a mathematically sophisticated cognitive flexibility assay with real-time measures of body metabolism and large-scale neural recordings. Parallel Aim 2 measures the metabolic costs of cognitive flexibility in a large sample of humans. It ensures the translational impact of our work while characterizing how humans differ in the energetic demands of cognitive flexibility. Finally, Aim 3 will determine if manipulating brain energy is sufficient to alter cognitive flexibility and related neural computations. It combines brief periods of fasting with targeted dietary supplementation in both species. Identifying the energetic costs and neural computations in cognitive flexibility will deliver insights that could protect flexibility in humans operating in energy-constrained environments. These insights could also pave the way towards next-generation, brain-inspired artificial intelligence that would be capable of operating flexibly within resource-limited environments.

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 06, 2025
Source ID
FA95502410305

Entities

People

  • Becket Ebitz

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • Université de Montréal

Tags

Readers

  • Distributed Systems and Data Platform Development
  • Neuroscience
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks